U.S. patent application number 12/907136 was filed with the patent office on 2011-08-04 for system for cardiac status determination.
This patent application is currently assigned to SIEMENS MEDICAL SOLUTIONS USA, INC.. Invention is credited to Prabhu Mukundhan, Hongxuan Zhang.
Application Number | 20110190643 12/907136 |
Document ID | / |
Family ID | 44342243 |
Filed Date | 2011-08-04 |
United States Patent
Application |
20110190643 |
Kind Code |
A1 |
Zhang; Hongxuan ; et
al. |
August 4, 2011 |
System for Cardiac Status Determination
Abstract
A system improves detection and diagnosis of blood pressure
based cardiac function and tissue activities by analyzing and
characterizing cardiac blood pressure signals (including
non-invasive and invasive blood pressure, discrete values and
continuous waveforms) using pressure signal variation and
variability calculation and evaluation. The system combines blood
pressure analysis with multi clinical related factors and
parameters to detect and quantify cardiac health status and
arrhythmia severity. The system determines an accurate time,
location and severity of cardiac pathology and events by
calculating blood pressure variability and statistical variation.
The accurately and reliably identifies cardiac disorders,
differentiates cardiac arrhythmias, characterizes pathological
severity, predicts the life-threatening events, and supports
evaluation of drug delivery effects.
Inventors: |
Zhang; Hongxuan; (Palatine,
IL) ; Mukundhan; Prabhu; (Chicago, IL) |
Assignee: |
SIEMENS MEDICAL SOLUTIONS USA,
INC.
Malvern
PA
|
Family ID: |
44342243 |
Appl. No.: |
12/907136 |
Filed: |
October 19, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61301338 |
Feb 4, 2010 |
|
|
|
Current U.S.
Class: |
600/486 ;
600/485 |
Current CPC
Class: |
A61B 5/0215 20130101;
A61B 5/021 20130101 |
Class at
Publication: |
600/486 ;
600/485 |
International
Class: |
A61B 5/0215 20060101
A61B005/0215; A61B 5/021 20060101 A61B005/021 |
Claims
1. A system for heart performance characterization and abnormality
detection, comprising: an interface for receiving a set of sampled
data values representing patient blood pressure occurring during
individual heart cycles of a plurality of sequential heart cycles;
a signal processor for generating a mathematical distribution using
the received sampled data set and calculating, (a) a first value
substantially comprising an average of the sampled data set, (b) a
second value substantially comprising a distribution value at a
first deviation point from the average of the data set and (c) a
ratio of the first and second value as a representation of blood
pressure variation; and a comparator for comparing said ratio with
a threshold value to provide a comparison indicator; and a patient
monitor for in response to said comparison indicator indicating
said ratio exceeds the threshold value, generating an alert message
associated with the threshold.
2. A system according to claim 1, wherein said plurality of
sequential heart cycles are successive heart cycles.
3. A system according to claim 1, wherein said average of the
sampled data set comprises at least one of, (a) an arithmetic mean,
(b) a median and (c) a root mean square value.
4. A system according to claim 1, wherein said second value
substantially comprises a standard deviation value of the
dataset.
5. A system according to claim 1, wherein said signal processor
generates a third value substantially comprising a distribution
value at a second deviation point from the average of the data
set.
6. A system according to claim 5, wherein said signal processor
generates a ratio of a difference between the first and second
values and a difference between the second and third values as
representation of blood pressure variation.
7. A system according to claim 5, wherein said second deviation
point is substantially twice the deviation from the average of the
data set as the first deviation point.
8. A system according to claim 1, wherein said blood pressure
comprises at least one of, (a) an intra-cardiac blood pressure, (b)
a systolic blood pressure, (c) a diastolic blood pressure, (d) an
invasive blood pressure and (e) a non-invasive blood pressure.
9. A system according to claim 1, wherein said comparator
determines a comparison indicator indicating whether said ratio
lies in a predetermined value range and said patient monitor, in
response to said comparison indicator indicating said ratio lies in
a predetermined value range, generates an alert message associated
with the value range.
10. A system according to claim 1, wherein said threshold value is
derived from recorded blood pressure data for said patient.
11. A system according to claim 1, wherein said threshold value is
derived from recorded blood pressure data for a population of
patients having similar demographic characteristics including at
least two of, (a) age, (b) weight, (c) gender and (d) height, to
those of said patient.
12. A system according to claim 1, wherein said signal processor
dynamically adjusts said threshold value in response to a
determined blood pressure variation of said patient.
13. A system according to claim 1, wherein said mathematical
distribution comprises at least one of (a) a Gaussian distribution,
(b) a normal distribution, (c) an Autoregression distribution model
and (d) an Autoregression and Moving Average distribution
model.
14. A system according to claim 1, including a repository of
mapping information, associating ranges of the ratio with
corresponding medical conditions and said comparator compares said
ratio with said ranges to provide a comparison indicator
identifying a medical condition and said patient monitor generates
an alert message identifying said medical condition.
15. A system according to claim 14, wherein said predetermined
mapping information associates ranges of the ratio with particular
patient demographic characteristics and with corresponding medical
conditions and said system uses patient demographic data including
at least one of, age weight, gender and height in comparing the
ratio with said ranges and generating an alert message indicating a
potential medical condition.
16. A system according to claim 1, wherein said first value and
said second value comprise amplitude values.
17. A method employed by at least one processing device for heart
performance characterization and abnormality detection, comprising
the activities of: receiving a set of sampled data values
representing patient blood pressure occurring during individual
heart cycles of a plurality of sequential heart cycles; generating
a mathematical distribution using the received sampled data set and
calculating, (a) a first value substantially comprising a
distribution value at a first deviation point from the average of
the data set, (b) a second value substantially comprising a
distribution value at a first deviation point from the average of
the data set and (c) a ratio of the first and second value as a
representation of blood pressure variation; and comparing said
ratio with a threshold value to provide a comparison indicator; and
in response to said comparison indicator indicating said ratio
exceeds the threshold value, generating an alert message associated
with the threshold.
18. A method employed by at least one processing device for heart
performance characterization and abnormality detection, comprising
the activities of: receiving a set of sampled data values
representing patient blood pressure occurring during individual
heart cycles of a plurality of sequential heart cycles; generating
a spectrum from the received sampled data set and calculating, (a)
a first value substantially comprising a frequency spectrum value
at a first peak point, (b) a second value substantially comprising
a frequency spectrum value at a second peak point, and (c) a ratio
of the first and second value as a representation of blood pressure
variation; and comparing said ratio with a threshold value to
provide a comparison indicator; and in response to said comparison
indicator indicating said ratio exceeds the threshold value,
generating an alert message associated with the threshold.
Description
[0001] This is a non-provisional application of provisional
application Ser. No. 61/301,338 filed Feb. 4, 2010, by H. Zhang et
al.
FIELD OF THE INVENTION
[0002] This invention concerns a system for heart performance
characterization and abnormality detection, by analyzing sampled
data values representing patient blood pressure occurring during
individual heart cycles of multiple sequential heart cycles.
BACKGROUND OF THE INVENTION
[0003] Blood pressure (BP) measures force applied to the walls of
arteries as a heart pumps blood through the body. The pressure is
determined by the force and amount of blood pumped, and the size
and flexibility of the arteries. Systolic and diastolic arterial BP
is not static but undergoes natural variations from one heartbeat
to another and throughout the day (in a circadian rhythm). Blood
pressure variation and changes may be utilized for patient health
status monitoring. Some recent studies successfully use power
spectra (such as indicating Low or High Frequency bandwidth) to
analyze blood pressure variability for detecting cardiac
abnormality. However, blood pressure variability may be affected by
many factors, such as age, disease, breathing control, physical
conditions and neurological status.
[0004] There are two types of blood pressure: systolic and
diastolic. Systolic blood pressure comprises pressure of the blood
when the heart has imparted the maximum pressure. Diastolic blood
pressure is the pressure when the heart is in the resting phase.
Blood pressure (BP) is the pressure exerted by circulating blood on
the walls of blood vessels, and is a principal vital sign. During
each heartbeat, BP varies between a maximum (systolic) and a
minimum (diastolic) pressure. The mean BP decreases as the
circulating blood moves away from the heart through arteries and
has its greatest decrease in the small arteries and arterioles, and
continues to decrease as the blood moves through the capillaries
and back to the heart through veins. The systolic pressure and
diastolic pressure may show different kinds of variation and trends
for different cardiac events or arrhythmias and pressure mean value
may not detect cardiac conditions.
[0005] Noninvasive auscultatory and oscillometric measurements are
simpler and quicker than invasive measurements have virtually no
complications, and are less unpleasant and painful for the patient.
However, noninvasive methods may yield lower accuracy and small
systematic differences in numerical results. Non-invasive
measurement methods are more commonly used for routine examinations
and monitoring. Systolic and diastolic arterial BPs change in
response to stress, nutritional factors, drugs, disease, exercise,
and momentarily from standing up. Sometimes the variations are
large. Hypertension refers to arterial pressure being abnormally
high, as opposed to hypotension, when it is abnormally low. Along
with body temperature, respiratory rate, and pulse rate, BP
measurements are the most commonly measured physiological
parameters. However known pressure data analysis typically fails to
comprehensively extract pathology related pressure information and
exclude non-pathology data and noise. This results in a high rate
of false alarm in cardiac pathology detection.
[0006] Known blood pressure analysis usually tracks the absolute
value of systolic and diastolic blood pressure measurements and
mean calculations discretely (e.g., Measuring Non-invasive blood
pressure every 5 minutes), which may fail to extract sufficient
pathology and event information. Known pressure analysis may also
fail to exclude noise factors (non-related pressure variation
factors, such as respiration) which may distort detection and
characterization accuracy of cardiac events or arrhythmias.
Further, known blood pressure analysis usually does not
differentiate the results of different pressure analysis, such as
systolic, diastolic, EoS (end of systolic pressure), EoD (end of
diastolic pressure) and fail to bridge cardiac arrhythmia diagnosis
and status characterization with pressure calculation based
multi-parameter analysis. A system according to invention
principles addresses these deficiencies and related problems.
SUMMARY OF THE INVENTION
[0007] A system improves detection and diagnosis of blood pressure
based cardiac function and tissue activities by analyzing and
characterizing cardiac blood pressure signals (including
non-invasive and invasive blood pressure, discrete values and
continuous waveforms) involving determining pressure signal
variation and calculated parameter variability. A system for heart
performance characterization and abnormality detection includes an
interface for receiving a set of sampled data values representing
patient blood pressure occurring during individual heart cycles of
multiple sequential heart cycles. A signal processor generates a
mathematical distribution using the received sampled data set and
calculates, (a) a first value substantially comprising an average
of the sampled data set, (b) second value substantially comprising
a distribution value at a first deviation point from the average of
the data set and (c) a ratio of the first and second value as a
representation of blood pressure variation. A comparator compares
the ratio with a threshold value to provide a comparison indicator.
A patient monitor in response to the comparison indicator
indicating the ratio exceeds the threshold value, generates an
alert message associated with the threshold.
BRIEF DESCRIPTION OF THE DRAWING
[0008] FIG. 1 shows a system for heart performance characterization
and abnormality detection, according to invention principles.
[0009] FIG. 2 illustrates spectral analysis of a continuous
pressure waveform and data series and variation characterization,
according to invention principles.
[0010] FIG. 3 shows Gaussian modeling based blood pressure pattern
analysis and variability calculation, according to invention
principles.
[0011] FIG. 4 shows a flowchart of a process performed by the
system for blood pressure data based patient health status and
pathology detection and diagnosis, according to invention
principles.
[0012] FIG. 5 shows different anatomical positions used for
invasive or non-invasive methods for sensing hemodynamic and blood
pressure signals, according to invention principles.
[0013] FIG. 6 shows simulation of data variation for blood pressure
monitoring and diagnosis for patient health status determination
and tracking, according to invention principles.
[0014] FIG. 7 shows a flowchart of a process for adaptive filtering
of blood pressure signals for blood pressure data series variation
and variability analysis, according to invention principles.
[0015] FIG. 8 shows an artificial neural network (ANN) used for
heart performance characterization and abnormality detection,
according to invention principles.
[0016] FIG. 9 shows a flowchart of a process used by a system for
heart performance characterization and abnormality detection,
according to invention principles.
DETAILED DESCRIPTION OF THE INVENTION
[0017] A system improves detection and diagnosis of blood pressure
based cardiac function and tissue activities by analyzing and
characterizing cardiac blood pressure signals (including
non-invasive and invasive blood pressure, discrete values and
continuous waveforms) using pressure signal variation and
variability calculation. The system combines blood pressure
analysis with multi clinical related factors and parameters to
detect and quantify cardiac health status and arrhythmia severity.
The system determines an accurate time, location and severity of
cardiac pathology and events by calculating blood pressure
variability and statistical variation. The system identifies
cardiac disorders, differentiates cardiac arrhythmias,
characterizes pathological severity, predicts life-threatening
events, and supports evaluation of drug delivery effects.
[0018] The system performs blood pressure variation analysis
involving analyzing different pressure values, such as systolic,
diastolic, EoS (end of systolic pressure), EoD (end of diastolic
pressure) values, to provide information for cardiac status and
event detection and interpretation. The system employs different
kinds of factors (patient information, medical history) and
synchronizes information (respiration, ECG) to improve pressure
based pathology detection and characterization. An ANN (artificial
neural network), fuzzy model or expert system, may be used for data
combination and analysis for pathology diagnosis. Different types
of pressure value and variation in the values have different value
ranges and may reflect different kinds of pathologies. Diastolic
pressure (higher than 90 mm Hg) indicates hypertension, systolic
pressure (lower than 90 mm Hg) indicates low blood pressure, for
example. The blood pressure analysis may be utilized for earlier
detection, diagnosis and characterization of cardiac events and
arrhythmias.
[0019] Blood pressure data is often corrupted and the system
facilitates elimination of electrical noise, patient movement-noise
(respiration, physical movement). System pressure data variation
calculation is used for, non-invasive (NIBP) and invasive
(intra-cardiac and other least invasive measurement) blood pressure
analysis as well as analysis of pressure from different portions of
the body, such as heart, arteries, veins, or other parts of the
body. Different kinds of pressure variation are used to track
cardiac pathologies and events of patients. In combination with
other data (such as age, patient history, respiration, pulse rate,
ECG signal, SPO2 signal data), the system excludes non-pathological
factors from pressure analysis, which facilitates cardiac status
tracking, detection and characterization. Blood pressure is
continually changing depending on activity, temperature, diet,
emotional state, posture, physical state, and medication use and
these factors are accommodated by system analysis.
[0020] FIG. 1 shows system 10 for heart performance
characterization and abnormality detection. Hemodynamic signals,
such as invasive and non-invasive blood pressure waveforms and
related waveform calculations (such as dP/dt, differential of blood
pressure), are used by system 10 to diagnose, evaluate and
quantitatively characterize heart function, arrhythmias, and
patient health status. NIBP measurement of an arm is typically
utilized for patient health status detection and characterization,
especially in noisy cases (for example, in an ICD installation, ECG
signals are noisy and contain artifacts). However, NIBP data
provides systolic, diastolic, mean pressure values (discrete
numbers) at intervals of a couple minutes for example (typically 5
minutes). System 10 provides more specific and detailed blood
pressure analysis to continuously monitor and quantify the status
and health of medical patients. System 10 determines variation and
variability of blood pressure data for tracking abnormality of
cardiac health and arrhythmias. The system determines variation and
variability of blood pressure measurements and performs
multi-parameter analysis by using a combination of information for
patient status determination and prediction.
[0021] FIG. 1 shows system 10 for heart performance
characterization and abnormality detection. System 10 comprises at
least one computer system, workstation, server or other processing
device 30 including interface 12, repository 17, patient monitor
19, signal processor 15, comparator 20 and a user interface 26.
Interface 12 receives a set of sampled data values representing
blood pressure of patient 11 occurring during individual heart
cycles of multiple sequential heart cycles. Signal processor 15
generates a mathematical distribution (e.g., a Gaussian
distribution, a normal distribution, an Autoregression distribution
model and an Autoregression and Moving Average distribution model)
using the received sampled data set and calculates a first value,
second value and ratio of the first and second value. The first
value substantially comprises an average of the sampled data set,
the second value substantially comprises a distribution value at a
first deviation point from the average of the data set and the
ratio of the first and second value comprises a representation of
blood pressure variation. Comparator 20 compares the calculated
ratio with a threshold value to provide a comparison indicator.
Patient monitor 19, in response to the comparison indicator
indicating the ratio exceeds the threshold value, generates an
alert message associated with the threshold.
[0022] System 10 uses blood pressure measurements to track and
monitor patient vital signs and health status. For invasive blood
pressure, the systolic and diastolic pressure are acquired each
beat. While in NIBP measurement, the systolic and diastolic
pressures are obtained in 2 minute intervals, for example, since
air inflation and deflation takes time. So together with discrete
pressure measurement, the pressure variability and variation is
used to track a trend and detect patient pathologies and
events.
[0023] FIG. 2 illustrates spectral analysis of a continuous
pressure waveform and data series and variation characterization.
System 10 analyzes a blood pressure signal waveform of an
individual heart beat cycle (e.g., 203) to determine a spectrum and
pressure variation by averaging signal data of multiple heart beat
signals to improve noise immunity. Similarly, other types of
modeling methods and approaches may be used by system 10 for data
analysis. Discrete or continuous blood pressure measurements may be
modeled based on an AR (Autoregression) model, Autoregression and
Moving Average (ARMA) model, and a Gaussian distribution.
Definitions.
[0024] A discrete data series, such as of systolic and diastolic
pressures, A.sub.i is a blood pressure measurement data series
(discrete pressure measurements and data values), E( ) is the
expectation of the data series and .mu. is the standard deviation
of the data series, and the pressure measurement variation is,
pressure_variation Discrete = E ( A i ) .mu. . ##EQU00001##
[0025] A continuous data set, B.sub.i comprises sample data of a
continuous blood pressure waveform (such as in invasive pressure
measurement), f.sub.B is the spectrum of the blood pressure
waveform data (for each heart cycle), the local maximum peak values
of f.sub.B are A.sub.1, A.sub.2, A.sub.3, respectively (205 FIG. 2)
and continuous pressure variation is,
pressure_variation Continuous _ 1 = A 1 A 2 ##EQU00002##
pressure_variation Continuous _ 2 = A 1 - A 2 A 2 - A 3 ,
##EQU00002.2##
[0026] FIG. 3 illustrates Gaussian modeling based blood pressure
pattern analysis and variability calculation. System 10 uses blood
pressure data to generate a Gaussian modeling curve (e.g., curve
303) for blood pressure pattern analysis. System 10 (FIG. 1)
performs statistical modeling and data matching using a Gaussian
curve matching and modeling function, for example,
Gaussian normal distribution = 1 .delta. 2 .pi. - ( x - E ) 2 / ( 2
.delta. 2 ) ##EQU00003##
System 10 calculates modeling based blood pressure variability
using,
First order variability = A 0 A 1 ##EQU00004## Second order
variability = A 0 - A 1 A 1 - A 2 ##EQU00004.2##
In which, A0 is the curve magnitude at an expectation point, A1,
A2, . . . are the modeling curve magnitude at different deviation
(e.g., standard deviation) points, such as .delta., 2.delta., . . .
as illustrated in curve 303 of FIG. 3. Different curves 303, 305
and 307 represent different modeling results based on the cardiac
blood pressure measurements and reflects statistical pattern change
and variability of blood pressure data. Hence by calculating and
quantifying blood pressure variability, patient health status, such
as cardiac tissue and function abnormality, is detected and
characterized. System 10 performs high order statistical
variability analysis for blood pressure modeling and calculation by
comparison of variability data with an abnormality evaluation
threshold. In an example, blood pressure data in a heart beat
window is processed by system 10 to provide a Gaussian model in
which A0, A1, A2, . . . , represent different statistical values in
a modeling curve e.g., curve 303. In response to receiving blood
pressure data of a successive heart beat cycle, system 10
adaptively uses data of the new successive cycle to update the
model and curve. Thereby system 10 provides a different data set
for each different window for real time curve calculation and for
determination of A0.sub.i, A1.sub.i, A2.sub.i, . . . .
[0027] FIG. 4 shows a flowchart of a process performed by the
system for blood pressure data based patient health status and
pathology detection and diagnosis. System 10 in step 412 following
the start at step 411, generates a first Gaussian blood pressure
curve (e.g., curve 303 FIG. 3) from a first data set. System 10 in
step 415, generates a second blood pressure curve (e.g., curve 305
FIG. 3) from a second data set. Curve 305 indicates changes and
variation from curve 303 in the pressure data curve. System 10 in
step 417, generates a third blood pressure curve (e.g., curve 307
FIG. 3) from a third data set. Curve 307 indicates changes in the
variation from curve 303. A data set of blood pressure
measurements, a1-an, is mapped to Gaussian curve 303 and based on
the curve, A0, A1, A2 . . . An values are derived according to
standard deviation level, .delta., 2.delta., . . . .
[0028] In step 423, system 10 determines different modeling result
values of A0, A1, A2, etc., for curves 303, 305 and 307 that are
used to calculate deviation and pressure variability. In step 426,
system 10 determines different variability values of different
level as follows. Higher order variability functions can also
readily be derived in the series and used for the blood pressure
analysis.
First order variability = A 0 A 1 ##EQU00005## Second order
variability = A 0 - A 1 A 1 - A 2 ##EQU00005.2## Higher order
example = ( A 0 - A 1 ) - ( A 1 - A 2 ) ( A 1 - A 2 ) - ( A 2 - A 3
) . ##EQU00005.3##
[0029] System 10 also calculates values employing high order
statistical theories. System 10 maps a blood pressure input data
set to a curve as output such as curves as illustrated in FIG. 3
having a data distribution including A0, A1, A2 values. The
variation calculation based on A0, A1, A2, indicates blood pressure
changes so that if there is an unexpected change, the data modeling
curve changes and the curve model switches from curve 303 to 305,
for example. The calculated amplitude and deviation change values
indicate pressure status change and quantifies curve mode changes.
A user identifies differences from the curve changes and the
calculated different order variability values that numerically
indicates changes. In step 429, system 10 iteratively repeats steps
412 to 426 until a warning threshold is exceeded. The process of
FIG. 4 ends at step 431.
[0030] FIG. 5 shows different anatomical positions including, head,
neck, heart, arm, wrist, finger and leg positions, used for
invasive or non-invasive methods for sensing hemodynamic and blood
pressure signals. System 10 uses a selectable kind of invasive or
non-invasive blood pressure sensor or transducer employing
different kinds of data sensing and acquisition method based on the
anatomical position at which blood pressure is being measured. For
example, for a heart position, ultrasound and electromagnetic field
methods may be used whereas lasers may be utilized in signal
acquisition and monitoring for a finger position. For wrist and arm
positions, different methods are selected based on measuring
convenience, including ultrasound and vibration sensing based
hemodynamic signal monitoring. Pressure at different sites may be
measured non-invasively or invasively. Intracranial blood pressure
data may be acquired in a head and neck, for example. System 10
calculates blood pressure variation values to facilitate earlier
detection of abnormality at different anatomical positions, such as
artery, vein and capillary positions to prevent fatal disease, such
as a brain hemorrhage. System 10 analyzes variation in acquired
pressure data (such as a systolic and diastolic value series) at
different anatomical positions to detect and characterize a blood
pressure pattern and determine a blood pressure trend to identify
patient pathologies and clinical events and identify location,
timing, severity and type of cardiac pathology. The analysis may
also involve statistical analysis such as a hypothesis test and
entropy analysis.
[0031] FIG. 6 shows simulation of blood pressure data variation for
patient health status determination and tracking. System 10 (FIG.
1) processes invasive blood pressure waveform 603 for variation
based blood pressure abnormality detection to derive pressure
variation waveform 605 and pressure variability first order
waveform 607 (using the functions previously described). System 10
determines abnormality at point 620 in the pressure variation
waveform 605 in response to the variation waveform exceeding
predetermined (+15% above normal range) threshold 610. System 10
determines abnormality at point 615 in the pressure variability
waveform 607 in response to the variation waveform exceeding
predetermined (+10% above normal range) threshold 612. In addition,
system 10 adaptively acquires blood pressure data at different
anatomical positions (including non-invasive and invasive
positions) for comparison to improve determining a location of an
abnormal vessel or tissue, for example. In certain diseases, one
heart chamber (such as a left ventricle) may not work at normal
squeezing speed, which may slow down a diastolic time period
(including time length, pressure amplitude, period between EoS (End
of Systolic) time to diastolic pressure time, rate of pressure
change dP/dt). System 10 detects these changes by analyzing a blood
pressure waveform.
[0032] System 10 blood pressure variation and variability analysis
adaptively and automatically selects a pressure parameter and
analysis function in response to data indicating patient medical
history and medical conditions. System 10 uses a single pressure or
combined pressures of different types or from different anatomical
sites in variation analysis. The pressure types include systolic
pressure, diastolic pressure, mean pressure, EOS (end of systolic)
pressure, EOD (end of diastolic) pressure, maximum blood pressure
and minimum blood pressure. System 10 adaptively selects a
calculation method such as a continuous or discrete variation
calculation method, first order or higher order variability
analysis, different models for data modeling (such as AR, ARMA,
Gaussian) in response to data indicating a clinical application,
for example. System 10 also enables a user to select a calculation
method for blood pressure variation analysis.
[0033] System 10 continuously monitors blood pressure in order to
detect and characterize changes in blood pressure for
differentiation of medical conditions. Blood pressures from
multiple different anatomical sites may be used for analysis and
the different pressures are dynamically selected for calculation in
response to a type of clinical application. In an example, a blood
pressure (systolic) value data set has 5 measurements (a data set
comprises a calculation window size of 5-10 heart cycles, for
example). The data set comprises, benign signals (reference values,
120, 116, 122, 123 and 128 and ongoing pressure measurement values
128, 137, 126, 111 and 109. System 10 calculates an average and
standard deviation value of average E=121.8, standard deviation
u=4.38 for benign signals and E=122.2, u=11.91 for ongoing signals.
The data indicates variation in the ongoing measurement and system
10 uses Gaussian modeling to derive data distribution related
parameters (using normalized amplitude in a distribution)
comprising A0=0.83, A1=0.32, A1=0.16 for the benign signal data set
and comprising A0=0.68, A1=0.43, A1=0.29 for the ongoing pressure
measurement data set.
[0034] System 10 uses the Gaussian modeled data to calculate first
order variability=2.6 and second order variability=3.2 for benign
signals and first order variability=1.58 and second order
variability=1.79 for ongoing signals. The variability calculation
indicates both first order and second order variation of the
ongoing signal have changed more than 20% from the corresponding
benign signal (pressure) data set variation. System 10 generates a
warning in response to systolic pressure variability exceeding a
predetermined threshold range (or a change in variation direction)
indicating a potential cardiac event or condition.
[0035] FIG. 7 shows a flowchart of a process for adaptive filtering
of blood pressure signals for blood pressure data series variation
and variability analysis. Blood pressure signal distortion may be
caused by medical device noise and patient noise (respiration and
movement), for example. In order to achieve better blood pressure
analysis, system 10 adaptively filters a blood pressure signal
based on patient context and noise for subsequent analysis to
determine blood pressure changes caused by drug delivery, clinical
events, and patient health status (e.g. cardiac tissue distortion,
function changes). System 10 adaptively selects a filtering method
in response to type of noise present such as electrical noise
needing 50/60 notch filtering, respiration needing high pass
filtering, patient movement needing singularity artifacts
suppression. After pre-filtering and data analysis, other filtering
may be used for singularity cleaning and shape filtering. The
system filters multiple hemodynamic signals, such as invasive blood
pressure signals from a blood pressure catheter, a NIBP signal, and
signals derived from other invasive or less invasive methods for
blood pressure acquisition.
[0036] System 10 in step 709 following summation filter stage 705,
segments and averages blood pressure signal data that was acquired
and digitized in step 703, performs variation and variability and
pattern analysis on one or more heart cycles and heart cycle
segments of the data and averages results over multiple cycles. The
resultant data is output following a further summation filter stage
715. System 10 in step 713 determines whether filtering is to be
applied at stage 705 and/or stage 715 based on noise content in
particular frequency ranges of the blood pressure signal such as
power line frequency and respiratory frequency determined by a
noise analysis performed in step 720. System 10 also determines the
type of filtering to be applied to the blood pressure signal based
on the analysis performed in step 720. The analysis of step 720
determines from other patient signals (such as a respiration
signal, vital sign signal), patient record information and data
indicating system and environment information what frequency ranges
need to be filtered with notch, low and high pass filters, for
example. The filtering is performed at one or more of summation
stages 705 and 715 using pre-data filter unit 711 and additional
filtering unit 717, respectively.
[0037] FIG. 8 shows an artificial neural network (ANN) system 807
used for heart performance characterization and abnormality
detection. ANN unit 807 employs blood pressure data series
variation analysis to identify cardiac disorders. ANN unit 807 maps
patient medical record data (age, treatment, medication) 820,
electrophysiological signal amplitude and frequency related
parameters and vital sign data (including respiration, ECG,
temperature and blood oxygen saturation) 823 and multi-channel
invasive and non-invasive blood pressure signal data values 826, to
output parameters 829. Output parameters 829 include blood pressure
variation values, a patient health status index and location, a
pathology severity indicator, a time of a cardiac event, a
pathology trend indication, a pathology type indication and
candidate treatment suggestions. ANN unit 807 structure comprises 3
layers, an input layer 810, hidden layer 812 and output layer 814.
ANN unit A.sub.ij weights are applied between input layer 810 and
hidden layer 812 components of the ANN computation and B.sub.pq
weights are applied between hidden layer 812 and calculation index
components 814 of the ANN computation. The A.sub.ij weights and
B.sub.pq weights are adaptively adjusted and tuned using a training
data set. ANN unit 807 incorporates a self-learning function that
processes signals 820, 823 and 826 to increase the accuracy of
calculated results.
[0038] ANN unit 807 maps input signals 820, 823 and 826 to a
candidate diagnosis or treatment suggestion 829 to localize a
tissue impairment within an organ and determine time of occurrence
within a heart cycle. ANN unit 807 also identifies arrhythmia type
(e.g., AF, MI, VT, VF), severity of arrhythmia treatment and
urgency level and is usable for automatic heart condition
detection, diagnosis, warning and treatment. Further unit 807
performs statistical analysis to construct a threshold used to
detect tissue impairment and diagnose and predict cardiac
arrhythmia and pathology.
[0039] Following a training phase with a training data set, ANN
unit 807 maps signals 820, 823 and 826 to data 829 indicating an
Arrhythmia type, Arrhythmia severity, candidate treatment
suggestions, localized tissue impairment information identifying
the cardiac arrhythmia position, pathology conducting sequence,
abnormal tissue area and focus of the disorder and irregularity,
for example. The severity threshold of a pathology mapping decision
may vary from person to person and is adjusted at the beginning of
analysis. The system may be advantageously utilized in general
patient monitoring and implantable cardiac devices for real time
automatic analysis and detection of cardiac arrhythmias and
abnormalities.
[0040] FIG. 9 shows a flowchart of a process used by system 10 for
heart performance characterization and abnormality detection. In
step 912 following the start at step 911, interface 12 receives a
set of sampled data values representing patient blood pressure
occurring during individual heart cycles of multiple sequential
(e.g., successive) heart cycles. The blood pressure comprises at
least one of, (a) an intra-cardiac blood pressure, (b) a systolic
blood pressure, (c) a diastolic blood pressure, (d) an invasive
blood pressure and (e) a non-invasive blood pressure. Signal
processor 15 in step 915 generates a mathematical distribution
using the received sampled data set and calculates, (a) a first
(e.g., amplitude) value substantially comprising a distribution
value at a first deviation point from the average of the data set,
(b) a second (e.g., amplitude) value substantially comprising a
distribution value at a first deviation point from the average of
the data set and (c) a ratio of the fast and second value as a
representation of blood pressure variation. In one embodiment, the
average of the sampled data set comprises one of, (a) an arithmetic
mean, (b) a median and (c) a root mean square value and the second
value substantially comprises a standard deviation value of the
dataset. The mathematical distribution comprises one of, (a) a
Gaussian distribution, (b) a normal distribution, (c) an
Autoregression distribution model and (d) an Autoregression and
Moving Average distribution model, for example. Further, in one
embodiment, the mathematical distribution comprises a spectrum, the
first value substantially comprises a frequency spectrum value at a
first peak point and the second value substantially comprises a
frequency spectrum value at a second peak point.
[0041] Signal processor 15 also generates a third value
substantially comprising a distribution value at a second deviation
point from the average of the data set. The second deviation point
is substantially twice the deviation from the average of the data
set as the first deviation point. Processor 15 generates a ratio of
a difference between the first and second values and a difference
between the second and third values as representation of blood
pressure variation.
[0042] In step 917 processor 15 stores in repository 17
predetermined mapping information associating predetermined ranges
of the ratio with corresponding medical conditions. The
predetermined mapping information associates ranges of the ratio
with particular patient demographic characteristics and with
corresponding medical conditions. Comparator 20 in step 923
compares the ratio with a threshold value and with the ranges to
provide a comparison indicator indicating a medical condition. The
system uses patient demographic data including at least one of, age
weight, gender and height in comparing the ratio with the ranges
and generating an alert message indicating a potential medical
condition. The threshold value is derived from recorded blood
pressure data for the patient or for a population of patients
having similar demographic characteristics including (a) age, (b)
weight, (c) gender and (d) height, to those of the patient. Signal
processor 15 dynamically adjusts the threshold value in response to
a determined blood pressure variation of the patient. In step 926,
in response to the comparison indicator indicating the ratio
exceeds the threshold value or lies within one of the ranges,
patient monitor 19 generates an alert message associated with the
threshold and identifying the medical condition. The process of
FIG. 9 terminates at step 931.
[0043] A processor as used herein is a device for executing
machine-readable instructions stored on a computer readable medium,
for performing tasks and may comprise any one or combination of,
hardware and firmware. A processor may also comprise memory storing
machine-readable instructions executable for performing tasks. A
processor acts upon information by manipulating, analyzing,
modifying, converting or transmitting information for use by an
executable procedure or an information device, and/or by routing
the information to an output device. A processor may use or
comprise the capabilities of a computer, controller or
microprocessor, for example, and is conditioned using executable
instructions to perform special purpose functions not performed by
a general purpose computer. A processor may be coupled
(electrically and/or as comprising executable components) with any
other processor enabling interaction and/or communication
there-between. A user interface processor or generator is a known
element comprising electronic circuitry or software or a
combination of both for generating display images or portions
thereof. A user interface comprises one or more display images
enabling user interaction with a processor or other device.
[0044] An executable application, as used herein, comprises code or
machine readable instructions for conditioning the processor to
implement predetermined functions, such as those of an operating
system, a context data acquisition system or other information
processing system, for example, in response to user command or
input. An executable procedure is a segment of code or machine
readable instruction, sub-routine, or other distinct section of
code or portion of an executable application for performing one or
more particular processes. These processes may include receiving
input data and/or parameters, performing operations on received
input data and/or performing functions in response to received
input parameters, and providing resulting output data and/or
parameters. A user interface (UI), as used herein, comprises one or
more display images, generated by a user interface processor and
enabling user interaction with a processor or other device and
associated data acquisition and processing functions.
[0045] The UI also includes an executable procedure or executable
application. The executable procedure or executable application
conditions the user interface processor to generate signals
representing the UI display images. These signals are supplied to a
display device which displays the image for viewing by the user.
The executable procedure or executable application further receives
signals from user input devices, such as a keyboard, mouse, light
pen, touch screen or any other means allowing a user to provide
data to a processor. The processor, under control of an executable
procedure or executable application, manipulates the UI display
images in response to signals received from the input devices. In
this way, the user interacts with the display image using the input
devices, enabling user interaction with the processor or other
device. The functions and process steps herein may be performed
automatically or wholly or partially in response to user command.
An activity (including a step) performed automatically is performed
in response to executable instruction or device operation without
user direct initiation of the activity.
[0046] The system and processes of FIGS. 1-9 are not exclusive.
Other systems, processes and menus may be derived in accordance
with the principles of the invention to accomplish the same
objectives. Although this invention has been described with
reference to particular embodiments, it is to be understood that
the embodiments and variations shown and described herein are for
illustration purposes only. Modifications to the current design may
be implemented by those skilled in the art, without departing from
the scope of the invention. The system performs blood pressure
variation analysis involving analyzing different pressure values to
provide information for cardiac status and event detection and
interpretation. Further, the processes and applications may, in
alternative embodiments, be located on one or more (e.g.,
distributed) processing devices on a network linking the units of
FIG. 1. Any of the functions and steps provided in FIGS. 1-9 may be
implemented in hardware, software or a combination of both.
* * * * *